Usefulness of combining inflammatory biomarkers and clinical scales in an emergency department to stratify risk in patients with infections

Objective. To determine whether combining biomarkers of inflammatory response and clinical scales can improve risk stratification of patients with suspected infection in a hospital emergency department (ED). Methods. Prospective observational study of ED patients treated for infections. We collected the following information on arrival: demographic and baseline clinical data, comorbidities, the focus of infection, and values for the following inflammatory biomarkers: leukocyte counts, and C-reactive protein, procalcitonin, and midregional proadrenomedullin (MR-proADM) concentrations. Scores on the following clinical scales were recorded based on the variables gathered: the SIRS (Systemic Inflammatory Response Syndrome) criteria, the qSOFA (Quick Sequential Organ Failure Assessment), and the NEWS (National Early Warning Score). The main outcome was a composite measure that included 30-day death or need for intensive care unit (ICU) admission. Results. A total of 473 patients with a mean (SD) age of 70.3 (19.2) years were included. The majority were men (257, 54.3%). Thirty-one (6.6%) died within 30 days and 16 (3.4%) were admitted to the ICU. The composite outcome (death or ICU admission) occurred in 45 patients (9.5%). The MR-proADM concentration, with an area under the receiver operating characteristic curve of 0.739 (95% CI, 0671-0.809) was a better predictor than the other biomarkers or clinical scales, although the differences between MR-proADM and either lactate concentration or the NEWS were not significant in the comparisons (P =.064). Combining the MR-proADM concentration with any of the scales did not significantly improve risk prediction. Conclusions. Risk stratification of patients with infection is a key part of ED decision-making. MR-proADM concentration is superior to other biomarkers and clinical prediction scales for predicting short-term prognosis in the ED. Combining MR-proADM measurement with other scales or measures does not improve the yield.